LidarView: The ParaView Lidar app
LidarView performs real-time visualization and processing of live captured 3D LiDAR data. LidarView is a ParaView-based application, which means it is based on the open source Paraview code-base, for loading data, showing it and interacting with it. On top of the ParaView infrastructure, LidarView provides specific realtime network data handling for multiple Lidar brands, as well as Lidar-related post-processing tools. It also has a specific User Interface, targeted at real-time viewing and analysis of Lidar data from major Lidar manufacturers.
LidarView is a brand-independent version of VeloView (that Kitware develops with Velodyne) that is independent of Velodyne and provides experimental algorithms and features under an Apache 2.0 licence. Contrasting with the brand-supported version, you may expect LidarView to crash in some situations.
The open source LidarView codebase serves as a base software that Kitware, or even other companies, can adapt and extend for custom needs such as specific point-cloud processing (SLAM, object detection, lane marking interpretation, roadsign recognition, point-cloud labelisation…).
LidarView can playback pre-recorded data stored in .pcap files, and can record live stream as .pcap file. LidarView have been tested showing more than 3 million points per second in realtime (VeloView with VLS-128/Alpha Puck sensor) or high framerate (OpsysView with 50 Lidar frames per second). LidarView provides tools to display, select and measure information about the points captured from the sensor. LidarView displays the distance measurements from the Lidar as point cloud data and supports custom color maps for multiple point properties such as intensity-of-return, time, distance, azimuth, dual return type, and laser id. The data can be exported as XYZ data in CSV format or screenshots of the currently displayed point cloud can be exported with the touch of a button. Export as LAS file is also possible if a localisation and orientation over time is available.
- Configurable listening ports
- Forward data received toward another port or/and ip address
- Direct georeferencing in LAS file using a time-synced GPS and IMU data file
- SLAM algorithm for mapping, without GPS/IMU. See example below.
- User configurable playback speed, and option to enforce playing speed by skipping frame if needed
- Underlying ParaView version is now version 5.4, providing many new tools
- Input from live sensor stream or recorded .pcap file
- Visualization of LiDAR returns in 3D + time including 3D position and attribute data such as timestamp, azimuth, laser id, etc
- Visualize the same data from multiple viewpoint, in multiple sub-windows, even using different colormap
- Spreadsheet inspector for LiDAR attributes
- Record to .pcap from sensor
- Export to CSV or VTK formats
- Record and export GPS and IMU data
- Ruler tool
- Visualize path of GPS data
- Show multiple frames of data simultaneously
- Show or hide a subset of lasers
Being mostly a showoff for innovative and experimental features, compilation from source is prefered following Developer Instructions.
Binary installers will be provided following our “How to Slam with LidarView” tutorial.
In order for sensor streaming to work properly, it is important to disable firewall restrictions for the Ethernet port. Disable the firewall completely for the ethernet device connected to the sensor or explicitly allow data from that Ethernet port of (including both public and private networks). When opening pre-recorded data or live sensor streaming data one is prompted to choose a calibration file. Please select the shipped file corresponding to your sensor model, or “Automatic from live stream”, if your sensor supports it.
The source code is available on : https://gitlab.kitware.com/LidarView/LidarView
Detailed instructions for building and packaging are available in developers guide the source code checkout under the Documentation folder.